Manager Call 2020-09-16
- Present: Judy, Laurent-Philippe, David, Huaiyu, Kimberly, Seth, Suzi, Pascale, Chris, Paul
- 1 Agenda
- 1.1 Projects updates
- 1.1.1 Anyone has blocking issues/delays/changes in project scope they need to report to managers /PIs?
- 1.1.2 GAF 2.2 (#4h) and GPAD/GPI 2.0 (#4i) (Kimberly)
- 1.1 Projects updates
Anyone has blocking issues/delays/changes in project scope they need to report to managers /PIs?
Nothing new to report at the moment.
GAF 2.2 (#4h) and GPAD/GPI 2.0 (#4i) (Kimberly)
- When do we want to make an announcement to the user community?
- This seems dependent upon a firm date for switchover - do we have that?
- We should also supply a test file for developers and include its location in the announcement.
- MGI and WB are nearly ready with their files; based on 2020-09-15 annotation call, other groups are coming along
- Propose a meeting between David, Kimberly, Helen, Penelope, and Alex to discuss UniProt files wrt timing of new file release and relations used
- Expect that we will need to continue to provide support (questions, clarifications) to groups as we get closer to the switchover
- specifications are out, waiting for feedback. WormBase and MGI nearly ready.
- out by end of year or early next year
- possibly start an announcement and supply a test file (available in the GH ticket)
- GAF 2.2 will be required, GPAD/GPI2.0 seems more flexible
- can we make this transition even simpler ?
- we can try to provide a GAF 2.0 to 2.2 converter (Pascale: I thought 2.2 -> 2.0 converter)
Action items: Kimberly
- Draft the announcement and provide a link to the test file
- Mention we'll provide a 2.0 -> 2.2 converter
- Goal: December for consortium and March release for external users
Noctua Imports (#3a)
- MGI/Wormbase imports
- Dustin has a new import on his USC server where David and Kimberly can check models for a select number (15-20) of genes, but code is still using GPAD 1.2 for import. Upshot: this will allow us to check the current state of the import rules for any errors, but does not reflect what the final import will include, e.g. annotation metadata.
- Not yet ingesting GPAD2.0. Metadata is not yet ingested (eg title, production/dev, created_by, modified_by, etc) and it still need some more discussions.
- The conversion into GO-CAMs seems to be near completion. Still work to be completed on the pipeline to complete this work.
- Minerva and GPAD output
- We need to schedule a virtual working meeting to focus on the GO-CAM GPAD output. Proposal: David and Kimberly create an agenda so we can schedule a meeting knowing what specific work needs to get done, and then find a good time for everyone. This meeting will required Ben, Jim, and Seth. Also Eric? Dustin?
- Once we do the import, how do we handle subsequent modifications of these annotations ? Ideally they would be done with Noctua.
Action items: Kimberly and David:
- finalize annotation properties specs GPAD2.0 (metadata/history)
- produce examples of issues with GPAD output
- draft an agenda for a working meeting (~2days) to defines the requirements for GPAD output
- priorities those issues to have different milestones and a better vision on when we could work with those ? (eg when for a 80% or 90% solution)
Next steps: 1. New MGI/WB files to be QC'ed (David & Kimberly) 2. Finalize history/annotation properties GPAD2.0 specs 3. Spec out GPAD output from GO-CAM models
Legacy data (#6)
Pascale, Laurent-Philippe Discussed offline with Paul, Laurent-Philippe, Seth and Pascale
- 1. gp2protein, gp2RNA files, gpi files: Any preferences for keeping/dropping them ?
- 2. File names: what about users citing old filenames ? see https://github.com/geneontology/archive-reconstruction/issues/1#issuecomment-680121448
- 3. Folder structure: can we avoid repetition in folder names ? see https://github.com/geneontology/archive-reconstruction/issues/1#issuecomment-680122454 . How do we plan for future changes ? For instance if annotation files are created for species rather than groups (this would also be a breaking change for Zenodo).
- 4. File browser: use the AWS S3 of the archive (https://geneontology-tmp.s3.amazonaws.com/index.html) for both current & release.geneontology.org ?
- 5. Do we store this archive only on release.geneontology.org or do we store it also on Zenodo *now* ? Pros: Zenodo is here for "always" and free, AWS S3 could be turned off (eg if cost not paid); Cons: everything in Zenodo would have a DOI and it could easily discourage any changes to the archive folder hierarchy / file names (situation we are in)
- ui implementations
- bug fixes
- Ongoing; Tremayne is submitting some PRs to noctua-dev for testing
- Adding: connect to existing nodes; re-use of fields in a given model
- Tremayne has a prototype that David and Kimberly can test. The prototype has much of the desired functionality but is not yet employing the necessary functionality from minerva which will be developed later when Ben is available to work on that.
- bug fixes
- testing and release framework